Re: OOPS when copying data from local to an external drive (ieee1394)

From: Zwane Mwaikambo
Date: Mon Mar 08 2004 - 01:54:58 EST


On Sun, 7 Mar 2004, Dmitry Torokhov wrote:

> I started getting oopses when cpying data from local IDE to an external
> Firewire drive. Not always, but quite often. The kernel is a bk pull a
> day before 2.6.4-rc2 was released, I do not see any ieee1394 updates
> since.
>
> Unfortunately the oops was not saves in the logs, so here is what I managed
> to write down:

> Oops: 00002 [#1]
> PREEMPT
> CPU: 0
> EIP: 0060 [<c0243d087>] Tainted: P
> EFLAGS: 00010047
> EIP is at hpsb_packet_sent+0x86/0x90
> eax: 00100100 ebx: dfd74000 ecx: dd6edfb0 edx: 00200200

A spot of linked list corruption.

> esi: 00000001 edi: dd6cdf60 ebp: c03e3ee0 esp: c03c3edc
> ds: 007b es: 007b ss: 0068
> Process swapper (pid: 0; threadinfo=c03c2000, task=c034a800)
> ....
> Call trace:
> [<co25306e>] dma_trm_tasklet+0xae/0x1b0

Does this patch help any?

Index: linux-2.6.4-rc1-mm2/drivers/ieee1394/ieee1394_core.c
===================================================================
RCS file: /home/cvsroot/linux-2.6.4-rc1-mm2/drivers/ieee1394/ieee1394_core.c,v
retrieving revision 1.1.1.1
diff -u -p -B -r1.1.1.1 ieee1394_core.c
--- linux-2.6.4-rc1-mm2/drivers/ieee1394/ieee1394_core.c 4 Mar 2004 04:12:44 -0000 1.1.1.1
+++ linux-2.6.4-rc1-mm2/drivers/ieee1394/ieee1394_core.c 8 Mar 2004 06:47:04 -0000
@@ -403,6 +403,8 @@ void hpsb_selfid_complete(struct hpsb_ho
void hpsb_packet_sent(struct hpsb_host *host, struct hpsb_packet *packet,
int ackcode)
{
+ unsigned long flags;
+
packet->ack_code = ackcode;

if (packet->no_waiter) {
@@ -413,7 +415,9 @@ void hpsb_packet_sent(struct hpsb_host *

if (ackcode != ACK_PENDING || !packet->expect_response) {
atomic_dec(&packet->refcnt);
+ spin_lock_irqsave(&host->pending_pkt_lock, flags);
list_del(&packet->list);
+ spin_unlock_irqrestore(&host->pending_pkt_lock, flags);
packet->state = hpsb_complete;
queue_packet_complete(packet);
return;
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/